fit_pipeline Module

  • Reference

Code used to fit pipeline.

Functions

edge_helper 

edge_helper(source_node_id: str | None, source_node_name: str | None, source_name: str | None, target_name: str | None, dst_node_id: str | None, dst_node_name: str | None, graph_json_dict: Dict[str, Any]) -> None

Parameters

source_node_id
source_node_name
source_name
target_name
dst_node_id
dst_node_name
graph_json_dict

fit_pipeline

Run a single iteration of an AutoML experiment.

This method is automatically called during a regular AutoML experiment. fit_pipeline will evaluate the pipeline for this iteration, fit the pipeline with the provided data, calculate the various metrics relevant for this experiment, and log all the results in the specified AzureML Run's history.

fit_pipeline(automl_pipeline: AutoMLPipeline, control_settings: ExperimentControlSettings, resource_settings: ExperimentResourceSettings, orchestration_settings: ExperimentOrchestrationSettings, pipeline_selection_settings: PipelineSelectionSettings, data_settings: ExperimentDataSettings, automl_run_context: AutoMLAbstractRunContext, elapsed_time: int | None = None, onnx_cvt: OnnxConverter | None = None) -> FitOutput

Parameters

automl_pipeline
Required

AutoMLPipeline object containing pipeline id and serialized script.

fit_pipeline_params
Required

User settings for fitting this pipeline

automl_run_context
Required

child run context object

elapsed_time
default value: None

How long this experiment has already taken in minutes

onnx_cvt
default value: None

The onnx converter.

control_settings
resource_settings
orchestration_settings
pipeline_selection_settings
data_settings

Returns

AzureML Run Properties for this child run